Most tech startups on Day 0: we don't have time for things like testing, architecture, and design. Just make something that works
Too many startups on Day N: why doesn't anything work anymore?
I'm not saying that the Day 0 approach was wrong on Day 0, but there comes a time where that has to change. If you wait too long to realize it, it may be too late to fix all the issues you allowed to happen.

